Abstract
Noble metals supported on reducible oxides, like CoOx and TiOx, exhibit superior activity in many chemical reactions, but the origin of the increased activity is not well understood. To answer this question we studied thin films of CoOx supported on an Au(111) single crystal surface as a model for the CO oxidation reaction. We show that three reaction regimes exist in response to chemical and topographic restructuring of the CoOx catalyst as a function of reactant gas phase CO/O2 stoichiometry and temperature. Under oxygen-lean conditions and moderate temperatures (≤150 °C), partially oxidized films (CoOx<1) containing Co0 were found to be efficient catalysts. In contrast, stoichiometric CoO films containing only Co2+ form carbonates in the presence of CO that poison the reaction below 300 °C. Under oxygen-rich conditions a more oxidized catalyst phase (CoOx>1) forms containing Co3+ species that are effective in a wide temperature range. Resonant photoemission spectroscopy (ResPES) revealed the unique role of Co3+ sites in catalyzing the CO oxidation. Density function theory (DFT) calculations provided deeper insights into the pathway and free energy barriers for the reactions on these oxide phases. These findings in this work highlight the versatility of catalysts and their evolution to form different active phases, both topological and chemically, in response to reaction conditions exposing a new paradigm in the catalyst structure during operation.
Similar content being viewed by others
Introduction
Catalysts are defined as materials that facilitate chemical reactions by providing special sites where reactants and products bind, react, and desorb with low energy barriers separating these steps. Although the composition and structure of the catalyst is usually assumed to be unaltered during reaction, many catalysts restructure by displacement of its atoms in response to the adsorption of reactants1,2,3. Here we demonstrate that in addition to topographic restructuring, chemical restructuring can also occur, adding an additional paradigm in the understanding of the working catalysts. In this work we illustrate this by showing that CoOx catalysts for the CO oxidation reaction undergo both chemical and topological changes in response to the reactant gas CO/O2 stoichiometry, evolving in three regimes characterized by different Co oxidation states. Partially oxidized films of cobalt (CoOx<1) deposited on Au containing Co0 were found to efficiently catalyze the CO oxidation reaction under oxygen-lean conditions and at temperatures lower than 150 °C. With increasing O2 content CoO forms first, which reacts with CO to form carbonates that poison the reaction for temperatures below 300 °C. Finally, under oxygen-rich conditions, more oxidized phases (CoOx>1) containing Co3+ species form that are more effective catalysts in a broad temperature range. Ambient Pressure X-ray photoelectron spectroscopy (APXPS) was employed to follow the catalyst oxidation state, the adsorption of CO and reaction products. While the various Co oxidation states can be monitored by APXPS during reaction, a precise identification and quantification is challenging due to the strong overlap of their 2p core level peaks. We overcame this difficulty using Resonant Photoelectron Spectroscopy (ResPES), which allowed us to precisely identify each Co oxidation state and relative concentration. DFT calculations provided in-depth perspectives related with CO oxidation reaction pathway and energy barriers on each CoOx phase.
Results and discussion
Deposition, oxidation, and wetting of cobalt films on Au(111)
Co2p XP spectra from a 1 MLE Co film on Au(111) before and after oxidation are shown in Fig. 1a. The bottom spectrum (black trace) displays the result before oxidation, showing the Co 2p3/2 level peak at 778.2 eV characteristic of metallic Co. After exposing the film to 10−6 Torr of O2 gas at RT for 60 s (=60 Langmuir units), a partially oxidized Co film was formed. This is shown in the red spectrum by the additional peak at 780.2 eV, strongly overlapping with the Co0 peak, and its satellite at 786.6 eV, both characteristic of Co2+. Fitting the two overlapping 2p3/2 peaks (Fig. S1), we estimate that ~25% of the Co atoms are oxidized to Co2+. We will refer to this film as CoO0.25. Annealing the film in 1 × 10−6 Torr O2 at 200 °C for 10 min led to the formation of CoO, with the peak at 780.2 eV from Co2+ now being dominant (Fig. 1a, blue trace). Figure 1b shows the corresponding O 1 s XPS region, with the lattice oxygen peak at 529.6 eV, for both CoO0.25 and CoO. The peak at 531.2 eV is due to adsorbed OH and CO from residual background H2O and CO gases4. Fig. 1c shows XPS of the Au 4 f and Co3p region after each of these treatments. With an incident photon energy of 260 eV, the kinetic energy of photoelectrons exiting from the Au surface is ~180 eV, with an inelastic mean free path of ~ 5 Å5 i.e., ~ 2 atomic layers. After deposition of 1 MLE of metallic Co (black trace in Fig. 1c), the Au 4 f peak intensity decreased by ~ 40% compared to the pristine Au(111) (gray trace in Fig. 1c). This attenuation is consistent with the double-layer island structure of metallic cobalt (see SI) as described by Morgenstern et al.6. After exposing this film to 1 × 10−6 Torr O2 and annealing to 200 °C (blue trace in Fig. 1c), the Au peak intensity decreased to ~ 20% of its clean surface value. This attenuation is consistent with the spreading of CoOx and can be described well by a layer-by-layer growth (Figs. S2 and S3).
Further oxidation at 200 °C in a 105 mTorr of an O2-rich reaction mixture (O2/CO = 20:1) completed the oxidation of cobalt to Co2+ and Co3+, characterized by peaks at 780.2 and 779.8 eV respectively (green trace in Fig. 1a). The Co3+ is further characterized by the increased intensity of the satellite at 789 eV. We will refer to this film as Co3O47. The presence of the 2+ and 3+ oxidation states of Co and their proportion in the film will be confirmed and quantified later using ResPES.
CO adsorption on CoOx<1 and on CoOx=1
Figure 2a shows the C 1 s XPS region from the CoO0.25 film before and after introduction of 100 mTorr of CO. Before CO introduction (bottom gray trace), the sample shows a C 1 s peak at 284 eV, due to adventitious C contamination4. In the presence of 100 mTorr of CO, a strong peak at 286 eV appears due to chemisorbed molecular CO (red trace), which adsorbs only on metallic Co0 sites8,9. The small peak at 289 eV is due to carbonate species formed on the CoO areas occupying ~25% of the oxide film area. A peak from gas-phase CO, with its fine vibrational structure is visible at 291.6 eV. After pumping out the CO gas, the COads peak nearly vanished (top gray trace), due to desorption to equilibrate with the reduced gas pressure. By contrast, on a CoO film under 100 mTorr CO at RT only the peak at 289 eV associated with carbonate species is observed (Fig. 2b, middle curve), along with the peak of gas-phase CO at 292 eV, but no molecularly adsorbed CO is observed. The area of the carbonate peak on the CoO film became about 4 times larger than that on the CoO0.25 film. The amount of carbonates is substantially increased by exposing the sample to CO2 instead of CO, as shown in the top trace in Fig. 2b.
CO oxidation reaction catalyzed by partially oxidized cobalt (CoOx<1)
In the previous section we identified the species formed by CO adsorption on CoOx<1 and CoOx=1 at RT. Here we follow the evolution of the CoO0.25 surface during the CO oxidation via the Mars van-Krevelen mechanism as a function of temperature. The surface composition, followed by APXPS, is shown in Fig. 3. In UHV and in 100 mTorr of CO at RT the spectra in the C 1 s region are similar to those in Fig. 2. Heating to 100 °C caused a decrease in the intensity of the C 1 s peak at 286 eV from adsorbed CO due to the new equilibrium with the gas phase at the higher temperature (Fig. 3b). This is shown by the O lattice peak at 529.6 eV and the Co0 peak at 778 eV (Fig. 3a, c), which remained essentially unchanged at this temperature. The O 1 s peak at 531.8 eV, responding to overlapping O peaks from adsorbed CO and OH, decreased due to thermal desorption. The new small peak at 283.0 eV originates from cobalt carbide (CoCx), suggesting CO dissociation at Co0 sites at elevated temperature8. Raising the temperature to 150 °C increased the reaction rate, as shown by the increase of the Co0 peak at 778.2 eV, the decrease of Co 2p3/2 peak at 780.2 eV from Co2+ (Fig. 3a), and the decrease of the lattice oxygen peak at 529.6 eV. All these changes confirm the reduction of CoO0.25 by reaction between CO and lattice oxygen.
CO oxidation reaction on cobalt monoxide
After oxidation of 1 MLE of Co to form CoO, with all cobalt atoms in the Co2+ state, CO was introduced in the chamber to a pressure of 100 mTorr. The reaction was monitored by APXPS as a function of temperature, with the results shown in Fig. 4. The spectra show that the intensity of the Co2+ in Fig. 4a and the O peaks in Fig. 4c remained largely unchanged up to 300°C but dropped rapidly thereafter. The decrease of these peaks can be attributed to the combined oxide reduction upon decomposition of the carbonate, and cobalt monoxide (CoO) deweting, that creates CoO clusters and exposes more Au surface, as shown by the rapid increase of the Au 4 f peak intensity in Fig. 4d and illustrated schematically in the inset. The deweting is the result of the formation of unstable carbonates, and reveals a topographical restructuring accompanying the change in oxidation state phase, from CoOx=1 to CoOx<1. Details of the reaction are shown in SI (Figs. S4 and S5). Introduction of 100 mTorr O2 at ~350 °C caused the reappearance of the Co2p peak (SI, Fig. S6a), indicating no loss of Co by formation of Co-carbonyls. We discuss this reaction mechanism further in the theoretical studies in below Section 3.6.
CO oxidation catalyzed by CoOx>1
In the previous sections, the oxide film contained Coδ+ species with δ ≤ 2. However, under oxygen-rich reaction conditions, oxide phases including O-Co-O trilayers7,10 or Co3O4 phase11, containing Co3+ are present, which has been proposed by several authors to be the most active site for CO oxidation12,13,14,15. To our knowledge, however, this has not been proven spectroscopically in operando conditions. To ascertain this important point, it is necessary to unambiguously distinguish spectroscopically the Co oxidation states, Co3+ and Co2+, both involving different partially filled and empty Co d-levels. This can be done by Resonant Photoelectron Spectroscopy (ResPES), as proposed and demonstrated by several groups16,17. Briefly, ResPES is based on the photoemission of electrons from d-band states, enhanced by the resonant excitation of core-level electrons to empty d-states that decay by an Auger process of energy equal to the initial X-ray. This is illustrated on the left panel of Fig. 5 for one of the oxidation states18,19. As the electronic configuration differs between Co3+ ([Ar]3d6) and Co2+ ([Ar]3d7), different resonant excitation energies for 2p to 3d transitions exist for each species. Since the Auger-mediated emission and the direct photoemission are undistinguishable processes with identical final state, a strong resonant enhancement of the photoemission spectra is observed. The photon energy for the resonant excitation can be experimentally determined by collecting valence band photoemission spectra as a function of X-ray energy. A maximum emission will be obtained at the resonant energy.
The experimental determination of the resonant energies for 1 MLE of Co3O4, containing Co3+ and Co2+ on Au is shown in the heat maps (bright to dark for high and low intensity) of Fig.5b, c. For CoO, only Co2+ species are present, with a resonant photon energy of ħω = 779.8 eV, as shown in the heat map. The VB d-states of Co2+, with peaks at ~5 eV and ~10 eV, are strongly enhanced at this photon energy (green trace in (Fig. 5d). For ħω = 781.2 eV, slightly off resonance, the spectrum is similar as expected, but less intense (blue trace), and for ħω = 772 eV (far from resonance), the VB spectrum (light blue trace) is dominated by the Au substrate. For Co3O4 (Fig. 5c) the resonant photon energy for the Co3+ site is 781.2 eV, as shown by the maximum in the heat map. The VB spectrum at this photon energy shows several Co3+ d-band peaks: a sharp one at 1.0 eV, and others around 10 eV, and 5.0 eV (Fig. 5e). The VB of the Au substrate at approx. 1, 2.5, and 6 eV20.
The contribution from the Co3+ and Co2+ ions, D(Co3+) and D(Co2+) in Fig. 5e, can be quantified by the difference in peak intensities relative to the Co2+ and to the off-resonance spectra respectively (length of the arrow lines in Fig. 5e), and can be used to determine the relative concentration of these species because the resonant enhancement ratio (RER), D(Co3+)/D(Co2+), is directly proportional to the ratio of the concentrations16, N(Co3+)/N(Co2+), with a linear correction factor (y) that can be determined from the known ratio in stoichiometric Co3O4 films, through the equation:
The RER of stoichiometric Co3O4 is 0.67. Note that the contribution of the d levels from the Au substrate is subtracted to get the D(Co2+), while the contribution of both Co2+ and Au are subtracted at ~1 eV BE to get D(Co3+), as respectively indicated by two vertical arrows in Fig. 5e. Since the N(Co3+)/N(Co2+) value for stoichimetric Co3O4 (Co2O3-CoO) is 2, the linear factor y = 2.99. Therefore, we can determine the concentration ratio, N(Co3+)/N(Co2+) of nonstoichiometric CoOx film through the measurement of the RER if it contains both Co3+ and Co2+ sites.
The ResPES spectra in Fig. 6a–d demonstrate the role of Co3+ species in 0.5 MLE of CoOx>1 on Au(111) in the CO catalytic oxidation reaction. The ResPES in Fig. 6a corresponds to the initial film in UHV where the film is not completely oxidized (Fig. 6e, black curve). Exposure to 100 mTorr O2 at RT, increased the peak of Co3+ at 778.6 eV in the XPS (Fig. 6e, red curve), more clearly evidenced by the rise of the characteristic resonant peaks at 10 eV, 5.0 eV and 1.0 eV in Fig. 6b. The RER of this intermediate CoOx is 0.3 and, therefore its N(Co3+)/N(Co2+) ratio is nearly unity. The oxide intermediates remained structurally stable at RT when the reaction conditions were switched to 100 mTorr CO (Fig. S6b, c). The nonstoichiometric cobalt oxide was reduced to CoO when the reaction temperature was raised to ~100 °C, which is clearly seen in the ResPES of Fig. 6c. The reduction of CoOx at 100 °C indicates the high reactivity of the intermediate oxide phase. Importantly, adding 100 mTorr of O2 to the 100 mTorr of CO at 100°C regenerated the intermediate phase as indicated by the reappearance of the Co3+ resonant peak at ~1.0 eV (Fig. 6d) and the red shift of the Co 2p3/2 peak (Fig. 6e, green curve). Since this nonstoichiometric oxide phase is structurally stable under CO and O2 mixtures (CO/O2 = 1), it appears to be the most catalytically active phase for CO oxidation at mild temperatures in oxygen-rich conditions. From its RER value of 0.47, the N(Co3+)/N(Co2+) ratio of this active phase is 1.4, indicating a higher ratio of Co3+ species at 100 °C compared with that at RT. Furthermore, the stoichiometry parameter “x” in this intermediate CoOx is 1.29, namely CoO1.29 ≈ Co3O3.87, which is structurally equal to the Co3O4 phase. Co3+ species were also proposed to be the active sites in Co3O4 nanorods for the cryogenic CO oxidation reaction14. Recent surface science results have also confirmed the facile activation of molecular O2 to peroxide (O22-) and superoxide species (O2-) at the oxygen vacancies on the surface of Co3O4(100) model catalysts15. This suggests that the newly-formed Co3+ in the reactive environment, together with oxygen vacancies, is responsible for the enhanced catalytic reactivity of our Au(111)-supported nonstoichiometric CoOx monolayer films.
DFT calculations of the reaction between CO and CoOx/Au
To further understand the reactivity of CoOx structures on Au(111), DFT calculations of the reaction between CO and surface O were performed on CoOx<1/Au(111), CoO/Au(111), and CoOx>1/Au(111) using the calculated structural models in Fig. S7.
For CoOx<1/Au, like on CoOx<1/Pt4, CO adsorbs exergonically by -0.35 eV atop a Co atom not bound to O, in agreement with the presence of molecular CO detected by APXPS (Fig. S5). The adsorbed CO can react with two types of surface O located at either a Face Centered Cubic site (FCC) or a Hexagonal Close Packed site (HCP) site (with respect to surface Co). The reaction between adsorbed CO and surface O is endothermic, by 1.35 eV for O at the FCC site and 1.17 eV for O at the HCP site, which agrees with the required heating of the CoOx<1/Au film before reaction between adsorbed CO and surface O could take place. Next, the formation of carbonate groups on CoO/Au was investigated by DFT calculations. It was found to be unlikely for carbonate groups to form at the terraces of a stoichiometric bilayer CoO film, because the formation of CO2 was calculated to be just as exothermic as that of CO32- (Fig. S4), indicating that, if formed, they should be unstable unless under a high CO2 partial pressure. We tested the sensitivity of carbonate group formation to changes in Co-Co spacing by performing the same calculations over a CoO film with a narrower 3.00 Å Co-Co spacing, but the same trend was observed. Since the structure of bulk CoCO3 is known21, we examined the possibility that the formation of carbonates induced the restructuring of the oxide layer. The calculations indicate that a flat CoCO3 overlayer on Au(111) is unstable and weakly bound. Instead, the optimized structures were found to dewet and relax away from the Au surface (Fig. S8). This is supported by the results in Fig. 4d where the dewetting of the oxide upon heating above 150 °C and beyond is revealed by the rapid increase in the Au4f XPS peak. Thus, it is likely that the experimentally observed carbonate groups are located on a restructured CoO terrace7,22 or that they form at the edge of CoO islands which retain their 2D structure, as previously reported for CoO films on Pt(111)4. We note that the formation of carbonates on the CoO film with 3.00 Å Co-Co spacing also induced a dewetting reconstruction of interfacial Co, where the Co cation detaches from the Au substrate and moves above the surface-bound O (Fig. S4f). This restructuring also supports a more complex structural transformation of CoO upon the formation of carbonate groups.
Finally, to understand the superior reactivity of CO oxidation on CoOx>1/Au, DFT calculations were performed to obtain the free energy barriers of the CO-lattice O reaction (Figs. 7a and S9). Two possible structures for the CoOx>1/Au were evaluated: a Co3O4 film (Fig. 7) and a CoO2 film (Fig. S9). The structure of CoOx films supported on Au and Pt under oxidative conditions have been extensively characterized. For CoOx/Au, it has been proposed that the O-rich CoO2 phase is the most stable configuration under O27,10,23. Following our previous calculations of the structure of Co3O4(111) surfaces, a Co-poor and O-rich surface was chosen to simulate a Co3O4(111) film24,25. On this termination, only Co cations originally in bulk tetrahedral sites and O anions are exposed. To initiate the reaction, CO binds weakly on exposed Co. The adsorbed CO can react readily with O atoms bound to Co neighbors by crossing a 0.49 eV barrier. At 100 °C and 100 mTorr of CO, the net free energy barrier for CO-lattice -O reaction is 1.04 eV. We note that this barrier is lower than the CO-lattice O reaction energy over CoO0.5 (1.17 eV, Fig. S5b), making Co3O4 more reactive. Further, under steady-state CO oxidation, the sub-oxidized phase was not observed. On the other hand, the CoO2/Au(111) film appears less reactive than Co3O4 as a 1.24 eV initial barrier is required for the reaction between CO gas and lattice O (Fig. S9). Weakly bound CO2 produced by this exothermic reaction can either desorb or further react with lattice O by crossing a 0.04 eV barrier (Fig. 7b). The formation of the O vacancy is linked to the reduction of subsurface Co. Upon the formation of the O vacancy, the magnetic moments of the two subsurface Co atoms surrounding the O vacancy shifted from 1.85 and 1.84 μB to 1.90 μB, while the magnetic moment of the third subsurface Co changed from -0.09 to 2.59 μB, indicating reduction. Note that the electronic energy of reaction is −1.80 eV relative to CO gas, which is much more exothermic than the reaction over CoO (Fig. S4). The CO32- group formed in this step is a bidentate species interacting with adjacent exposed Co. Although easy to form, the CO32- group can also decompose by crossing a 0.71 eV barrier (Fig. 7b). Carbonates formed by reaction between CoO2 and the CO2 product are also easy to decompose, requiring an even smaller barrier of 0.21 eV (Fig. S9). Under a low CO2 partial pressure (PCO2 = 0.1 mTorr), the exergonic CO2 desorption prevents the easily formed CO32- from poisoning the surface. Although the Au substrate does not participate in the CO oxidation reaction because of the weak adsorption of CO and inefficient dissociation of O2, it can still modify the reactivity of the first CoOx layer in contact with the substrate. Our most recent results suggest that Au does indeed modify slightly the reactivity of the first monolayer CoO film, but has a lessened effect on the reactivity of the second layer, which remain slightly lower than that of the first.
In summary, the structure and reactions of Au-supported ultrathin CoOx catalyst films in the presence of CO/O2 mixtures, were studied with the goal of determining the evolution of the catalyst structure and the nature of the active sites involved in CO oxidation. It was found that CoOx<1 films have a high chemical reactivity towards CO oxidation under oxygen-lean conditions at mild temperatures (from 100 °C to 150 °C), which we attribute to the presence of O vacancies near Coδ+ where CO can adsorb and react with neighboring lattice oxygen. In contrast, on stoichiometric CoO, the CO adsorbs forming carbonates that poison the CO oxidation and trigger restructuring and dewetting of the oxide film. The formation of carbonates, the decomposition, and dewetting of the CoOx to form 3D clusters divide the reaction into two regimes of high catalytic activity. One is characteristic of oxygen-lean conditions containing both Co0 and Co2+, the other is characteristic of oxygen-rich conditions containing Co2+ and Co3+ species. The Co3+ can be reduced by CO to Co2+ and regenerated by O2 back to Co3+ at temperatures of 100 °C and below. Using ResPES, we demonstrated the important role of Co3+ as the most active catalyst site under reaction conditions. DFT calculations indicate that the high reactivity is due to a lower energy barrier for C-O bond formation and shows the pathway leading to restructuring and dewetting of the oxide film upon formation of carbonates. Our findings provide a general understanding of the enhanced catalytic reactivity of cobalt oxide catalysts and underline the double restructuring paradigm of chemical (i.e. Co oxidation state) and topographical (CoOx detachment and dewetting from the Au substrate) restructuring of the catalyst induced by the reactants composition and temperature.
Methods
Experimental methods
APXPS measurements were performed at beamline 9.3.2 of the Advanced Light Source (ALS) at the Lawrence Berkeley National Laboratory, at beamline 23-ID-2 (IOS) of the National Synchrotron Light Source II (NSLS-II) at Brookhaven National Laboratory, and at the Experimental Station 13-2 at the Stanford Synchrotron Radiation Light source (SSRL) at SLAC National Accelerator Laboratory. The Au(111) surface was cleaned by cycles of sputtering (5 min at 3 × 10−5 Torr of Ar+ at 1 keV energy), and annealing (10 min at 500 °C) until only Au was detected by XPS. Cobalt films were deposited on the clean Au(111) surface by evaporation from a Co rod (Goodfellow, 99.99 + %) using a SPECS e-beam evaporator. The amount of Co was measured by using XPS peak intensities, calibrated with a quartz crystal microbalance (QCM). The amount of Co on the surface is reported in monolayer equivalents (MLE), defined as the amount of Co that would form a complete monolayer if its wetting of Au was perfect. The reported MLE values have an estimated error bar of ±20% (see SI). The coverage of Co0 for a given number of MLE was determined using the Co 3p to Au 4 f intensity ratio (Fig. S2) in comparison with simulated ratios from the quantitative photoelectron simulation package SESSA v2.226 (Fig. S3). After deposition, the films were oxidized by exposure to O2 at room temperature (RT) or at 200 °C. To ensure high CO purity, the CO gas was passed through a carbonyl trap heated to ~240 °C before entering the measurement chamber. Total gas pressures were monitored with Baratron capacitance pressure gauges. Photon energies of 740, 475, 260, and 920 eV were used to generate photoelectrons with kinetic energies between 150 and 200 eV for the O 1 s, C 1 s, Co 3p, Co 2p, and Au 4f photoelectrons, respectively, which have mean free paths of ~ 5 Å. Reported binding energies are given with respect to the Fermi level.
Computational methods
DFT calculations were performed using the Vienna Ab initio Simulation Package (VASP) version 5.4.127,28,29. The exchange-correlation energy was calculated using the Perdew-Burke-Ernzerhof (PBE) functional30. Spin polarization was used in all calculations. The projector-augmented wave (PAW) method was used to describe the core electrons. The one-electron wavefunctions were expanded using a set of plane waves with kinetic energy up to 500 eV. To correct for the self-interaction error of Co 3d electrons, a Hubbard-like onsite repulsion term (DFT + U) was included in the calculations using Dudarev’s approach31, with an effective U value (Ueff) of 3.5 eV, which has been used in the literature to study the bulk and surface redox properties of CoOx32. The CoO row-wise antiferromagnetic state was maintained in all calculations. More details regarding the spin state of Co cations are provided in the SI4,33. Structural relaxation for reaction intermediates was performed using the conjugate gradient algorithm. Transition states were first searched using the nudged elastic band (NEB) and climbing image (CI) NEB algorithms34,35. The highest energy image of each CI-NEB calculation was then refined using the Dimer and quasi-Newton algorithms36. The electronic structure in each self-consistent field (SCF) cycle was considered converged when the difference in total energy of consecutive steps fell below 10-6 eV. Atomic positions were considered converged when the Hellmann-Feynman forces on unconstrained atoms fall below 0.02 eV/Å. Free energies of CO and CO2 gas were approximated using their translational and rotational partition functions37. The structure of ultrathin CoOx films on the hexagonal (111) surfaces of Au and Pt by STM has been studied extensively in the past6,10,22,38. In all cases the films form hexagonal lattices with slightly different unit cells from that of the metal substrate leading to formation of Moire patterns. Our previous STM and DFT studies of CoOx/Pt(111)4 and FeOx/Pt(111)39,40,41,42,43,44,45,46, and the more recent work of Zeuthen et al. on FeOx/Pd(111)47, again report similar structures. Here we follow the crystallographic structure of the CoOx/Au(111) films and their change during reaction by theoretical modeling using DFT. Details of the structural models are described in the SI (Fig. S7).
Data availability
All data are available upon request.
References
Wu, C. H. et al. Bimetallic synergy in cobalt–palladium nanocatalysts for CO oxidation. Nat. Catal. 2, 78–85 (2018).
Eren, B. et al. Activation of Cu(111) surface by decomposition into nanoclusters driven by CO adsorption. Science 351, 475–478 (2016).
Tao, F. et al. Break-up of stepped platinum catalyst surfaces by high CO coverage. Science 327, 850–853 (2010).
Kersell, H. et al. CO oxidation mechanisms on CoOx-Pt thin films. J. Am. Chem. Soc. 142, 8312–8322 (2020).
Powell, C. J. & Jablonski, A. NIST Electron Inelastic-Mean-Free-Path Database 71, Version 1.0. (1999).
Morgenstern, K. et al. Cobalt growth on two related close-packed noble metal surfaces. Surf. Sci. 601, 1967–1972 (2007).
Fester, J. et al. Phase transitions of cobalt oxide bilayers on Au(111) and Pt(111): the role of edge sites and substrate interactions. J. Phys. Chem. B 122, 561–571 (2018).
Wu, C. H. et al. Ambient-pressure X-ray photoelectron spectroscopy study of cobalt foil model catalyst under CO, H2, and their mixtures. ACS. Catalysis 7, 1150–1157 (2017).
Böller, B., Ehrensperger, M. & Wintterlin, J. In situ scanning tunneling microscopy of the dissociation of CO on Co(0001). ACS. Catalysis 5, 6802–6806 (2015).
Walton, A. S. et al. Interface controlled oxidation states in layered cobalt oxide nanoislands on gold. ACS Nano 9, 2445–2453 (2015).
Li, M. & Altman, E. I. Shape, morphology, and phase transitions during Co oxide growth on Au(111). The. J. Phys. Chem. C. 118, 12706–12716 (2014).
Wang, Y.-G., Yang, X.-F. & Li, J. Theoretical studies of CO oxidation with lattice oxygen on Co3O4 surfaces. Chin. J. Catal. 37, 193–198 (2016).
Wang, H.-F. et al. Origin of extraordinarily high catalytic activity of Co3O4 and its morphological chemistry for CO oxidation at low temperature. J. Catal. 296, 110–119 (2012).
Xie, X. et al. Low-temperature oxidation of CO catalysed by Co(3)O(4) nanorods. Nature 458, 746–749 (2009).
Liu, Y. et al. Surface oxygen vacancies on reduced Co3 O4 (100): superoxide formation and ultra-low-temperature CO oxidation. Angew. Chem. Int. Ed. Engl. 60, 16514–16520 (2021).
Lykhach, Y. et al. Quantitative analysis of the oxidation state of cobalt oxides by resonant photoemission spectroscopy. J. Phys. Chem. Lett. 10, 6129–6136 (2019).
Prince, K. C. et al. 2p resonant photoemission study ofTiO2s. Phys. Rev. B 55, 9520–9523 (1997).
Lykhach, Y. et al. Counting electrons on supported nanoparticles. Nat. Mater. 15, 284–288 (2016).
Wu, Q.-H. et al. Resonant photoemission spectroscopy study of electronic structure of V2O5. Chem. Phys. Lett. 430, 309–313 (2006).
Salmerón, M. et al. Photoelectron-spectroscopy study of the electronic structure of Au and Ag overlayers on Pt(100), Pt(111), and Pt(997) surfaces. Phys. Rev. B 28, 6758–6765 (1983).
Pertlik, F. Structures of hydrothermally synthesized cobalt(II) carbonate and nickel(II) carbonate. Acta Crystallogr. Sect. C Cryst. Struct. Commun. 42, 4–5 (1986).
Rattigan, E. et al. The cobalt oxidation state in preferential CO oxidation on CoOx/Pt(111) investigated by operando X-ray photoemission spectroscopy. Phys. Chem. Chem. Phys. 24, 9236–9246 (2022).
Fester, J. et al. Comparative analysis of cobalt oxide nanoisland stability and edge structures on three related noble metal surfaces: Au(111), Pt(111) and Ag(111). Top. Catal. 60, 503–512 (2016).
Yan, G., et al. Water on oxide surfaces: a triaqua surface coordination complex on Co3O4 (111). J. Am. Chem. Soc. 141, 5623-5627 (2019).
Yan, G. & Sautet, P. Surface structure of Co3O4 (111) under reactive gas-phase environments. ACS. Catalysis 9, 6380–6392 (2019).
Smekal, W., Werner, W. S. M. & Powell, C. J. Simulation of electron spectra for surface analysis (SESSA): a novel software tool for quantitative Auger-electron spectroscopy and X-ray photoelectron spectroscopy. Surf. Interface Anal. 37, 1059–1067 (2005).
Kresse, G. & Furthmuller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B Condens. Matter 54, 11169–11186 (1996).
Kresse, G. & Hafner, J. Ab initio molecular dynamics for liquid metals. Phys. Rev. B Condens. Matter 47, 558–561 (1993).
Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15–50 (1996).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
Dudarev, S. L. et al. Electron-energy-loss spectra and the structural stability of nickel oxide: an LSDA+U study. Phys. Rev. B 57, 1505–1509 (1998).
Bajdich, M. et al. Theoretical investigation of the activity of cobalt oxides for the electrochemical oxidation of water. J. Am. Chem. Soc. 135, 13521–13530 (2013).
Giordano, L. et al. Interplay between structural, magnetic, and electronic properties in aFeO∕Pt(111)ultrathin film. Phys. Rev. B. 76, 075416 (2007).
Henkelman, G. & Jónsson, H. Improved tangent estimate in the nudged elastic band method for finding minimum energy paths and saddle points. J. Chem. Phys. 113, 9978–9985 (2000).
Henkelman, G., Uberuaga, B. P. & Jónsson, H. A climbing image nudged elastic band method for finding saddle points and minimum energy paths. J. Chem. Phys. 113, 9901–9904 (2000).
Henkelman, G. & Jónsson, H. A dimer method for finding saddle points on high dimensional potential surfaces using only first derivatives. J. Chem. Phys. 111, 7010–7022 (1999).
Hill, T. L., An Introduction To Statistical Thermodynamics (Courier Corporation,1986).
Rattigan, E. et al. Dewetting transition of CoO/Pt(111) in CO oxidation conditions observed in situ by ambient pressure STM and XPS. J. Phys. Chem. C 127, 8547–8556 (2023).
Galloway, H. C., Benítez, J. J. & Salmeron, M. The structure of monolayer films of FeO on Pt(111). Surf. Sci. 298, 127–133 (1993).
Galloway, H. C., Benítez, J. J. & Salmeron, M. Growth of FeOx on Pt(111) studied by scanning tunneling microscopy. J. Vac. Sci. Technol. A: Vac. Surf. Films 12, 2302–2307 (1994).
Galloway, H. C., Sautet, P. & Salmeron, M. Structure and contrast in scanning tunneling microscopy of oxides: FeO monolayer on Pt(111). Phys. Rev. B Condens. Matter 54, R11145–R11148 (1996).
Kim, Y. J. et al. Interlayer interactions in epitaxial oxide growth: FeO on Pt(111). Phys. Rev. B 55, R13448–R13451 (1997).
Merte, L. R. et al. CO-induced embedding of Pt adatoms in a partially reduced FeO(x) film on Pt(111). J. Am. Chem. Soc. 133, 10692–10695 (2011).
Liu, Y. et al. Structure and electronic properties of interface-confined oxide nanostructures. ACS Nano 11, 11449–11458 (2017).
Chen, H. et al. Active phase of FeOx/Pt catalysts in low-temperature CO oxidation and preferential oxidation of CO reaction. J. Phys. Chem. C 121, 10398–10405 (2017).
Liu, Y. et al. Enhanced oxidation resistance of active nanostructures via dynamic size effect. Nat. Commun. 8, 14459 (2017).
Zeuthen, H. et al. Structure of stoichiometric and oxygen-rich ultrathin FeO(111) films grown on Pd(111). J. Phys. Chem. C 117, 15155–15163 (2013).
Acknowledgements
This work was supported by the Office of Basic Energy Sciences (BES), Chemical Sciences, Geosciences, and Biosciences Division, of the U.S. Department of Energy (DOE) under Contract DE-AC02-05CH11231, FWP CH030201 (Catalysis Research Program). L. J. F. acknowledges support from the Alexander von Humboldt Foundation, Bonn, Germany. It used resources of the Advanced Light Source, a U.S. DOE Office of Science User Facility under contract no. DE-AC02-05CH11231, the 23-ID-2 (IOS) beamline of the National Synchrotron Light Source II, a User Facility operated for the DOE Office of Science by Brookhaven National Laboratory under Contract No. DE-SC0012704, and the Stanford Synchrotron Radiation Light Source, SLAC National Accelerator Laboratory, supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515. X.Z. was supported by NSF-BSF grant number 1906014. The DFT calculations in this work used computational and storage services associated with the Hoffman2 cluster at the UCLA Institute for Digital Research and Education (IDRE), and the Bridges-2 cluster at the Pittsburgh Supercomputing Center (supported by National Science Foundation award number ACI-1928147) through the Extreme Science and Engineering Discovery Environment (supported by National Science Foundation grant number ACI-1548562) grant TG-CHE170060.
Author information
Authors and Affiliations
Contributions
M.S. oversaw the project. H.C., L.J.F., H.K., X.Z., and J.O.M. conduct the APXPS experiment supported by the beamline scientists: S.N., A.H., I.W., and H.O. Data analysis were done by H.C. and M.S. Simulations were carried out by G.Y. and P.S. The paper was written by H.C., G.Y., P.S., and M.S., with contributions from all authors. A.T.B. and M.J. helped to review and edit the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Chen, H., Falling, L.J., Kersell, H. et al. Elucidating the active phases of CoOx films on Au(111) in the CO oxidation reaction. Nat Commun 14, 6889 (2023). https://doi.org/10.1038/s41467-023-42301-7
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-023-42301-7
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.